Methods and apparatus of this type are known in practice. Usually methods and apparatus of this type use a bank of correlators which is employed in the receiver. On the basis of the correlation results with the received signal, which is performed in the bank of correlators, the symbol of the received signal can be detected. For this detection the output of each correlator is the input argument of a mathematical function. The mathematical function is maximum for the correlator corresponding to the maximum function. According to said method and apparatus, which uses a pre-determined set of symbols, the symbol detection can be performed in such a way that it minimizes sensitivity to noise in the received signal. Examples of possible pre-determined sets of symbols are given in the high speed standard in the IEEE 802.11b standard. This IEEE 802.11b standard is especially meant for the 2.4 [GHz] band, also called Industrial Scientific Medical Band (ISMB-band), in the United States. It is noticed that corresponding bands are available in most other regions in the world. Users have free access to the ISMB-band if they comply with the standards of the standard proposal. A first important issue of the standard proposal is that the each used symbol has a relatively flat frequency power spectrum, which minimises risks of jamming fellow-users. A second important issue of the IEEE 802.11b standard is that not all possible symbols in the pre-determined set of symbols are used. This results in a redundant and robust detection mechanism.
A first disadvantage of the known methods and apparatus for the detection of a symbol of a received signal is the large number of correlators which are used in the correlator-bank. A second disadvantage of the type of known methods and apparatus is the large processing power which is required for performing the mathematical function, which function is used for the selection of a correlator in the correlator-bank, and operates on the complex output of each of the correlators. This mathematical function normally calculates the length of complex input argument, which leads to at least two multiplications per correlator. In order to reduce the processing power several approximations of said mathematical function have been proposed. However, these approximations only yields sub-optimal detection performance.